Something that provides some amount of protection is the idea of ethical injunctions.
âEthical injunctions are rules not to do something even when itâs the right thing to do. (That is, you refrain âeven when your brain has computed itâs the right thing to doâ, but this will just seem like âthe right thing to doâ.)â
The problem is, this doesnât help with some of these examples. Ethical injunctions would protect us against the first item. A 1940â˛s EA might say âYes, eugenics is a good thing, see the literature, but we still shouldnât kill or sterilise people without their consent to make it happen.â
Lobotomies are a bit harder, but I think a suitably strong ethical injunction of âDonât perform surgeries without peopleâs consentâ would protect us there too. What it wouldnât protect us from would be thinking lobotomies were a good thing and managing to convince patients to let us do them.
Finally, ethical injunctions gives us no protection against recovered memory therapy. If you grant as a prior that RMT works, there is no harm being doneâyou are simply bringing old memories to light and bringing guilty people to justice.
So thereâs basically two separate problems hereâhow should EA avoid being on the wrong side of history morally, and how should EA avoid being on the wrong side of history scientifically?
I think our best bet for the first is to hold to universal moral principles like âThou shalt not killâ and avoid allowing clever arguments to convince us to do things for the greater good. I think EA already does this pretty wellâthereâs a strong norm against, say, lying about our top charitiesâ effectiveness to solicit donations.
For the second one...I donât think thereâs a way out of that one. Science is humanityâs best guess at the time, so we canât do better except in areas we might happen to have a comparative advantage inâpsychology isnât one of them. All we can do is be open to changing our minds when the evidence stacks up against $WRONG_THEORY, so we should continue to promote epistemic norms of reasoning transparency and soliciting criticism of the movement.
Something that provides some amount of protection is the idea of ethical injunctions.
âEthical injunctions are rules not to do something even when itâs the right thing to do. (That is, you refrain âeven when your brain has computed itâs the right thing to doâ, but this will just seem like âthe right thing to doâ.)â
The problem is, this doesnât help with some of these examples. Ethical injunctions would protect us against the first item. A 1940â˛s EA might say âYes, eugenics is a good thing, see the literature, but we still shouldnât kill or sterilise people without their consent to make it happen.â
Lobotomies are a bit harder, but I think a suitably strong ethical injunction of âDonât perform surgeries without peopleâs consentâ would protect us there too. What it wouldnât protect us from would be thinking lobotomies were a good thing and managing to convince patients to let us do them.
Finally, ethical injunctions gives us no protection against recovered memory therapy. If you grant as a prior that RMT works, there is no harm being doneâyou are simply bringing old memories to light and bringing guilty people to justice.
So thereâs basically two separate problems hereâhow should EA avoid being on the wrong side of history morally, and how should EA avoid being on the wrong side of history scientifically?
I think our best bet for the first is to hold to universal moral principles like âThou shalt not killâ and avoid allowing clever arguments to convince us to do things for the greater good. I think EA already does this pretty wellâthereâs a strong norm against, say, lying about our top charitiesâ effectiveness to solicit donations.
For the second one...I donât think thereâs a way out of that one. Science is humanityâs best guess at the time, so we canât do better except in areas we might happen to have a comparative advantage inâpsychology isnât one of them. All we can do is be open to changing our minds when the evidence stacks up against $WRONG_THEORY, so we should continue to promote epistemic norms of reasoning transparency and soliciting criticism of the movement.